Xbox 360 Beta Kits in Action

[SanGreal]

Again, as I said in the post, in certain conditions, its quite easily conceivable for that to be the case, because those conditionas are non-optimal for SLI but will work better on a higher performance single board.

I only asked for clarifaction because in your original post you said "quite easy for G70 to be faster than 6800 SLI." compared to "it was more than twice as fast" which are vey different statements! I appreciate the response.

His comment just seems like a sweeping generality, in my opinion.

 

[Shifty Geezer]

Yes different teams come up with different artwork.. but those 2 games will still be the most comparible games first gen and also (arguably..) the most interesting. We will see which system can handle the best physics and AI.

Different teams doesn't just mean different artworks, but different coding abilities too. Not all devlopers are alike in ability.

compare DnD:Heroes on XB with CON on PS2. They are alike games - the same genre, the same style visuals. CON on PS2 looks much better. Yet it wouldn't be right to say therefore PS2>XB, as in the case it's the devs of CON>the devs of DnD:H.

The same CON engine was used for the cross platform title 'The Bard's Tale' and there was to my knowledge no obvious difference between platforms. I think the XB incarnation had higher-res textures, but they were pretty darned good anyway on PS2.

The only way to compare performance is like development. In essence it'll be towards the middle/end of the console's lives whre we see the best of the best of all genres, and then we'll see if one platform consistantly outdoes the other.

If you're looking for an earlier indication I would look to EA. EA aren't reknowned for relly pushing the consoles, and they only deal with cross-platform. Comparing EA titles to date has not shown XB>PS2 to a large degree mostly because they don't push either system. Now to date EA have made a few comments on how fanatastic Cell/PS3 is. If this is true, and if they really do find that and aren't just talking nonsense or only comparing alpha kits instead of final hardware, then it raises the possibility that because they like it so much they make more of the hardware. If there's no obvious difference on EA games between platforms, it might not show that the systems are comparable as they may just be targetting the lowest common denominator. But if one system is visibly better than the other, it IMO would show that system is intrinsically more powerful for the simple reason EA won't be doing anything clever to exploit the system's intricate advantages. This is likely to be PS3 going from EA's comments, though as may be the case, perhaps they were raving about Cell because they had Cell to play with, and didn't rave about XeCPU because it just wasn't available then?

But in the main, comparing different games from different devs isn't going to show anything worthwhile about the platform's relative capabilities.

 

[3roxor]

Different teams doesn't just mean different artworks, but different coding abilities too. Not all devlopers are alike in ability.

Yes yes I know but the best devs using the Ps1 won't be able to create anything looking nearly as good compared to some average Xbox game.

Now to date EA have made a few comments on how fanatastic Cell/PS3 is. If this is true, and if they really do find that and aren't just talking nonsense

Knowing EA..

But in the main, comparing different games from different devs isn't going to show anything worthwhile about the platform's relative capabilities.

Come on.. Look at 30 comparible Xbox/Ps2 games made by 30 different developers and you can most definately conclude which system is better at what. I do agree that these first gen games won't stretch both systems so it might be too early to make solid conclusions.

 

[Shifty Geezer]

The only way to compare performance is like development. In essence it'll be towards the middle/end of the console's lives whre we see the best of the best of all genres, and then we'll see if one platform consistantly outdoes the other.

 

[Dave Baumann]

I only asked for clarifaction because in your original post you said "quite easy for G70 to be faster than 6800 SLI." compared to "it was more than twice as fast" which are vey different statements! I appreciate the response.

<sigh>

Here is the post again:

If you have a crap load of offscreen / RTT effects then its quite easy for G70 to be faster than 6800 SLI.[/url]

As has already Bjorn picked up, the speed was not the operative part of that post, but the bolded element. If some here don't understand the significance of that then I would urge them to fully read and understand our SLI article.

His comment just seems like a sweeping generality, in my opinion.

Again, Bjorn has picked it up - his comment is going to be very focused on what Epic are doing, but his comment is given without bounds so people freely interpret in the manner they please. His comment is very specific to the work they are doing on UnrealEngine3 and the games they have in development which are also very specific to the particular usage scenarios in that.

 

[Qroach]

In other words. G70 is NOT faster than SLI, unless your running a bunch of render to texture effects off screen?

 

[Farid]

compare DnD:Heroes on XB with CON on PS2. They are alike games - the same genre, the same style visuals.

Those two game have another point in common, they use the same engine.
The CoN version is the better version, though, that's sure.

 

[Shifty Geezer]

No they don't. The SB team had no hand in DND:H, not even licensing their older BG: DA engine. The other games to use the CON engine are...

Bard's Tale
Combat Elite: WWII Paratroopers
Fallout Brotherhood of Steel

Team17 were working on a remake of AlienBreed using the engine too IIRC but it never saw the light of day. Shame as it was ideally suited IMO, and the engine could've handled the dark, moody style+lots of fancy lighting as the original worked towards.

 

[Farid]

No they don't. The SB team had no hand in DND:H, not even licensing their older BG: DA engine. The other games to use the CON engine are...

My bad.
Since you're the specialist for anything Snowblind Engine related, I won't even check Google.

What I don't know is why I was certain that the game used the Snowblind Engine...

 

[Powderkeg]

Here is the post again:

If you have a crap load of offscreen / RTT effects then its quite easy for G70 to be faster than 6800 SLI.

As has already Bjorn picked up, the speed was not the operative part of that post, but the bolded element. If some here don't understand the significance of that then I would urge them to.

It's quite easy for the G70 to be faster than 6800 SLI in any case. SLI does not double performance nor power. At best, it offers about a 30% boost over a single card, and that's nothing more than a memory bandwidth increase.

In SLI, shader instructions have to be duplicated on both cards, so it is in effect, a single GPU. Each card has to render the exact same frame, with the exact same effects. The only advantage they have over a single card is that they render less of the frame, which is like lowering the resolution to increase your frame rate.


If the G70 wasn't more powerful than 6800 SLI, I would be very very worried.

 

[Bjorn]

In SLI, shader instructions have to be duplicated on both cards, so it is in effect, a single GPU. Each card has to render the exact same frame, with the exact same effects. The only advantage they have over a single card is that they render less of the frame, which is like lowering the resolution to increase your frame rate.

Your whole post is way of base but i can't for the life of me figure out how you can say that the only advantage they have over a single card is that they (each card in the SLi configuration) render less of the frame and then not think that this means that it's more powerful.

 

[blakjedi]

I soo want to see the difference between HeavenlySword and Ninety-Nine Nights because it will say alot about the differences between Ps3 and Xbox360..

That doesnt make sense. If you have either title on both platforms that would say alot about the systems, if the game was coded by the same team. Your highlighted measure is meaningless as it exists...

 

[Titanio]

That doesnt make sense. If you have either title on both platforms that would say alot about the systems, if the game was coded by the same team. Your highlighted measure is meaningless as it exists...

An equally valid perspective is that multiplatform titles rarely take full advantage of one particular system..one system may get "dragged down" by the other etc. and thus multiplatform titles are an ineffective measure of a system's capability, or of highlighting real differences.

Directly comparing different exclusives that push each system may have its own problems, but that said I'm looking to the "prestige" exclusives from the big guns (technically) to demonstrate the capability of each system (and yeah, I may draw comparison using them).

 

[blakjedi]

That doesnt make sense. If you have either title on both platforms that would say alot about the systems, if the game was coded by the same team. Your highlighted measure is meaningless as it exists...

An equally valid perspective is that multiplatform titles rarely take full advantage of one particular system..one system may get "dragged down" by the other etc. and thus multiplatform titles are an ineffective measure of a system's capability, or of highlighting real differences.

Directly comparing different exclusives that push each system may have its own problems, but that said I'm looking to the "prestige" exclusives from the big guns (technically) to demonstrate the capability of each system (and yeah, I may draw comparison using them).

Fair to say. Id say if you compared say Splinter Cell on both systems and saw overwhelming differences for one system or another THATS a good indicator only because UbiSoft as a company does tend to push the capabilities of the systems they work on .... SC3 being a great example.

 

[SanGreal]

<sigh>...

Again, I was only asking for clarification on your opinion as to the degree that you would consider the g70 more powerful under those conditions. You provided the clarification, which I am greatful for. You don't have to keep repeating your justification as I understood it the first time . I apologize if it somehow came accross that I was debating the validity of your comments.

 

[Powderkeg]

Your whole post is way of base but i can't for the life of me figure out how you can say that the only advantage they have over a single card is that they (each card in the SLi configuration) render less of the frame and then not think that this means that it's more powerful.

Was what I said too complicated for you?

Both cards render the same frame. Therefore, the GPU on both cards are doing the exact same thing.

You cannot double the shader instructions just because you run SLI, for instance. You have to run the same instructions on both cards, effectively making it perform like a single GPU.


For example:

SLI 6800 GT only gives a 17 FPS (17%) frame rate increase over a single 6800 GT, and only a 7 FPS improvement over a single overclocked card.

In fact, a single overclocked 6800 GT is faster than SLI 6800's.


The only advantage SLI has is in memory bandwidth, it doesn't offer a significant increase in shader instruction processing. The G70 does offer a major increase in shader instruction processing power, so it would easily be more powerful than SLI 6800's.

 

[Gholbine]

If the games are not designed to work with SLI or Xfire in mind, then you won't see much of a performance boost. However if they are, you'll certainly get more than a "17%" performance increase.

 

[Powderkeg]

If the games are not designed to work with SLI or Xfire in mind, then you won't see much of a performance boost. However if they are, you'll certainly get more than a "17%" performance increase.

Show me.


Last time I checked, SLI works at the driver level, and doesn't require the games to be coded for it.


And it's the same for Far Cry above, Doom 3:

Halo

Lock On

And in some cases, a single card can actually outperform SLI, such as in Half Life.

SLI is not significantly more powerful than a single card. 10-30% performance increases are the norm, and it is only when you go to extremely high resolutions with large amounts of FSAA and AF applied do you see any significant advantage.

 

[Arty]

To counter those graphs: 6800 Ultra vs 6800 Ultra SLI

http://firingsquad.com/hardware/nvidia_geforce_7800_gtx/images/sc1600.gif (78% increase)

http://firingsquad.com/hardware/nvidia_geforce_7800_gtx/images/d3h1600.gif (67% increase)

http://firingsquad.com/hardware/nvidia_geforce_7800_gtx/images/bf21280.gif (74% increase)

Mr.Baumann said that the G70 is faster than 6800 SLI under certain conditions, he did avoid saying that G70 is faster than 6800 Ultra SLIs, hence maybe not twice as fast as a 6800 Ultra.

 

[west]

If the games are not designed to work with SLI or Xfire in mind, then you won't see much of a performance boost. However if they are, you'll certainly get more than a "17%" performance increase.

There are no games now which are designed with SLI in mind. There will be no games that are designed with SLI in mind. So lets deal with the 10-30% increase the SLI does provide.